This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
Make \x{...} consistently produce UTF-8.
authorSimon Cozens <simon@netthink.co.uk>
Sun, 29 Oct 2000 19:36:48 +0000 (19:36 +0000)
committerJarkko Hietaniemi <jhi@iki.fi>
Sun, 29 Oct 2000 21:00:47 +0000 (21:00 +0000)
Subject: Re: \x{...} is confused
Message-ID: <20001029193648.A6287@pembro4.pmb.ox.ac.uk>

p4raw-id: //depot/perl@7485

t/pragma/utf8.t
toke.c

index 768da05..93a5bc4 100755 (executable)
@@ -10,7 +10,7 @@ BEGIN {
     }
 }
 
-print "1..181\n";
+print "1..191\n";
 
 my $test = 1;
 
@@ -326,11 +326,16 @@ sub nok_bytes {
 {
     # bug id 20001009.001
 
-    my($a,$b);
-    { use bytes; $a = "\xc3\xa4"; }  
-    { use utf8;  $b = "\xe4"; }
-    { use bytes; ok_bytes $a, $b; $test++; } # 69
-    { use utf8;  nok      $a, $b; $test++; } # 70
+    my ($a, $b);
+
+    { use bytes; $a = "\xc3\xa4" }
+    { use utf8;  $b = "\xe4"     } # \xXX must not produce UTF-8
+
+    print "not " if $a eq $b;
+    print "ok $test\n"; $test++;
+
+    { use utf8; print "not " if $a eq $b; }
+    print "ok $test\n"; $test++;
 }
 
 {
@@ -726,3 +731,72 @@ __EOMK__
     }
 }
 
+{
+    # tests 182..191
+
+    {
+       my $a = "\x{41}";
+
+       print "not " unless length($a) == 1;
+       print "ok $test\n";
+       $test++;
+
+       use bytes;
+       print "not " unless $a eq "\x41" && length($a) == 1;
+       print "ok $test\n";
+       $test++;
+    }
+
+    {
+       my $a = "\x{80}";
+
+       print "not " unless length($a) == 1;
+       print "ok $test\n";
+       $test++;
+
+       use bytes;
+       print "not " unless $a eq "\xc2\x80" && length($a) == 2;
+       print "ok $test\n";
+       $test++;
+    }
+
+    {
+       my $a = "\x{100}";
+
+       print "not " unless length($a) == 1;
+       print "ok $test\n";
+       $test++;
+
+       use bytes;
+       print "not " unless $a eq "\xc4\x80" && length($a) == 2;
+       print "ok $test\n";
+       $test++;
+    }
+
+    {
+       my $a = "\x{100}\x{80}";
+
+       print "not " unless length($a) == 2;
+       print "ok $test\n";
+       $test++;
+
+       use bytes;
+       print "not " unless $a eq "\xc4\x80\xc2\x80" && length($a) == 4;
+       print "ok $test\n";
+       $test++;
+    }
+
+    {
+       my $a = "\x{80}\x{100}";
+
+       print "not " unless length($a) == 2;
+       print "ok $test\n";
+       $test++;
+
+       use bytes;
+       print "not " unless $a eq "\xc2\x80\xc4\x80" && length($a) == 4;
+       print "ok $test\n";
+       $test++;
+    }
+}
+
diff --git a/toke.c b/toke.c
index b007de4..274e506 100644 (file)
--- a/toke.c
+++ b/toke.c
@@ -1187,13 +1187,13 @@ S_scan_const(pTHX_ char *start)
     register char *d = SvPVX(sv);              /* destination for copies */
     bool dorange = FALSE;                      /* are we in a translit range? */
     bool didrange = FALSE;                     /* did we just finish a range? */
-    bool has_utf = FALSE;                      /* embedded \x{} */
+    bool has_utf8 = FALSE;                     /* embedded \x{} */
     UV uv;
 
     I32 utf = (PL_lex_inwhat == OP_TRANS && PL_sublex_info.sub_op)
        ? (PL_sublex_info.sub_op->op_private & (OPpTRANS_FROM_UTF|OPpTRANS_TO_UTF))
        : UTF;
-    I32 thisutf = (PL_lex_inwhat == OP_TRANS && PL_sublex_info.sub_op)
+    I32 this_utf8 = (PL_lex_inwhat == OP_TRANS && PL_sublex_info.sub_op)
        ? (PL_sublex_info.sub_op->op_private & (PL_lex_repl ?
                                                OPpTRANS_FROM_UTF : OPpTRANS_TO_UTF))
        : UTF;
@@ -1327,7 +1327,7 @@ S_scan_const(pTHX_ char *start)
 
        /* (now in tr/// code again) */
 
-       if (*s & 0x80 && thisutf) {
+       if (*s & 0x80 && this_utf8) {
            STRLEN len;
            UV uv;
 
@@ -1343,7 +1343,7 @@ S_scan_const(pTHX_ char *start)
                while (len--)
                    *d++ = *s++;
            }
-           has_utf = TRUE;
+           has_utf8 = TRUE;
            continue;
        }
 
@@ -1416,9 +1416,10 @@ S_scan_const(pTHX_ char *start)
                        yyerror("Missing right brace on \\x{}");
                        e = s;
                    }
-                   {
+                   else {
                        STRLEN len = 1;         /* allow underscores */
                        uv = (UV)scan_hex(s + 1, e - s - 1, &len);
+                       has_utf8 = TRUE;
                    }
                    s = e + 1;
                }
@@ -1435,8 +1436,8 @@ S_scan_const(pTHX_ char *start)
                 * There will always enough room in sv since such escapes will
                 * be longer than any utf8 sequence they can end up as
                 */
-               if (uv > 127) {
-                   if (!thisutf && !has_utf && uv > 255) {
+               if (uv > 127 || has_utf8) {
+                   if (!this_utf8 && !has_utf8 && uv > 255) {
                        /* might need to recode whatever we have accumulated so far
                         * if it contains any hibit chars
                         */
@@ -1468,9 +1469,9 @@ S_scan_const(pTHX_ char *start)
                         }
                     }
 
-                    if (thisutf || uv > 255) {
+                    if (has_utf8 || uv > 255) {
                        d = (char*)uv_to_utf8((U8*)d, uv);
-                       has_utf = TRUE;
+                       this_utf8 = TRUE;
                     }
                    else {
                        *d++ = (char)uv;
@@ -1499,7 +1500,7 @@ S_scan_const(pTHX_ char *start)
                    res = new_constant( Nullch, 0, "charnames", 
                                        res, Nullsv, "\\N{...}" );
                    str = SvPV(res,len);
-                   if (!has_utf && SvUTF8(res)) {
+                   if (!has_utf8 && SvUTF8(res)) {
                        char *ostart = SvPVX(sv);
                        SvCUR_set(sv, d - ostart);
                        SvPOK_on(sv);
@@ -1508,7 +1509,7 @@ S_scan_const(pTHX_ char *start)
                        /* this just broke our allocation above... */
                        SvGROW(sv, send - start);
                        d = SvPVX(sv) + SvCUR(sv);
-                       has_utf = TRUE;
+                       has_utf8 = TRUE;
                    }
                    if (len > e - s + 4) {
                        char *odest = SvPVX(sv);
@@ -1587,7 +1588,7 @@ S_scan_const(pTHX_ char *start)
     *d = '\0';
     SvCUR_set(sv, d - SvPVX(sv));
     SvPOK_on(sv);
-    if (has_utf)
+    if (has_utf8)
        SvUTF8_on(sv);
 
     /* shrink the sv if we allocated more than we used */
@@ -6553,7 +6554,7 @@ S_scan_str(pTHX_ char *start, int keep_quoted, int keep_delims)
     register char term;                        /* terminating character */
     register char *to;                 /* current position in the sv's data */
     I32 brackets = 1;                  /* bracket nesting level */
-    bool has_utf = FALSE;              /* is there any utf8 content? */
+    bool has_utf8 = FALSE;             /* is there any utf8 content? */
 
     /* skip space before the delimiter */
     if (isSPACE(*s))
@@ -6565,7 +6566,7 @@ S_scan_str(pTHX_ char *start, int keep_quoted, int keep_delims)
     /* after skipping whitespace, the next character is the terminator */
     term = *s;
     if ((term & 0x80) && UTF)
-       has_utf = TRUE;
+       has_utf8 = TRUE;
 
     /* mark where we are */
     PL_multi_start = CopLINE(PL_curcop);
@@ -6611,8 +6612,8 @@ S_scan_str(pTHX_ char *start, int keep_quoted, int keep_delims)
                   have found the terminator */
                else if (*s == term)
                    break;
-               else if (!has_utf && (*s & 0x80) && UTF)
-                   has_utf = TRUE;
+               else if (!has_utf8 && (*s & 0x80) && UTF)
+                   has_utf8 = TRUE;
                *to = *s;
            }
        }
@@ -6640,8 +6641,8 @@ S_scan_str(pTHX_ char *start, int keep_quoted, int keep_delims)
                    break;
                else if (*s == PL_multi_open)
                    brackets++;
-               else if (!has_utf && (*s & 0x80) && UTF)
-                   has_utf = TRUE;
+               else if (!has_utf8 && (*s & 0x80) && UTF)
+                   has_utf8 = TRUE;
                *to = *s;
            }
        }
@@ -6701,7 +6702,7 @@ S_scan_str(pTHX_ char *start, int keep_quoted, int keep_delims)
 
     if (keep_delims)
        sv_catpvn(sv, s, 1);
-    if (has_utf)
+    if (has_utf8)
        SvUTF8_on(sv);
     PL_multi_end = CopLINE(PL_curcop);
     s++;